Anal. Chem. 1983,55, 163-166
Table 11. Kinetic Parameters of the (Hg)Zn Mi) Couple at an HMDE in 1.0 M)/Zn2+ Molar Supporting Electrolytea
-i o ' , A/cmz
k " , cm/s
supporting electrolyte
-___
this work
this
work
KNO,
2.3
K C1
10-3 2.6 X
X
10-3
lit.
3.5 X
4.5
X
5.0
X
lit. 6.3 X 1 0 - ~ c
b
4.0 X 10-3b
Refer-
a Measured at 50 Hz and at -1.00 V (vs. SCE). ence 14. Reference 16.
2 ,--
54
v
% -0.2 -c4
-c.6 -3 a
-1.9
-$
2
-$.4
-I 6
POTER' I A L ,1 vs %E Figure 4. Potential dependlence of the capacity component of adenine in a borate buffer pH 9 at a DME. Adenine concentration was as (A) 0.0; (B) 0.031 mM; (C) 0 083 mM; (D) 0.125 mM; (E) 0.25 mM; (F) 0.5 mM; (G)1.0 mM; (H) 2.0 mM; (I) 4.0 mM; (J) 5.0 mM; (K) 8.0 mM. Curves were taken at 100 Hz; ac current was 0.5 pA and scan rate was 5 mV/s.
quency and changing potential. The first system selected was that of the Zn(II)/Zn(Hg) reaction in aqueous KNOBor KC1 media (Figure 3). This reaction has been investigated by using ac impedance methods. Results from the literature and our expermental data under the same solution condition are listed in Table 11. Both ',i and k" were calculated according to literature procedures (1, 14). Given the uncertainties in solution purification and the usual systematic errors from laboratory to laboratory, the values obtained by our instrumentation appear to be acceptable. More importantly, however, is the fact that a typical impedance diagram used for the calculation of id (or k o ) can be made in 5 min as compared to 1 h or more needed for manual determination. In Figure 4 a plot of the impedance of double layer capacity c vs. potential is shown for adenine at different concentrations. Measurements were dame at a DME in borate buffer solution (ionic strength 0.5 M) using sampled measurements. The adsorption of adenine on Hg has been studied before using a quadrature component of ac sinusoidal polarography (15).
163
The data obtained in the present work suggest that using impedance measurements one can obtain a Zc-V curve in approximately the same period of time as in ac polarography but with better accuracy. The accuracy of measurement of Zc by the described ac impedance method is about 0.5% in comparison to the 1-2% accuracy in quadrature ac polarography. In conclusion the instrument described in this paper offers advantages not only in cost but in a substantial savings of time. If the frequency is changed automatically, numerous complex plane spectra may be rapidly obtained in the time that is usually needed for the obtaining of a single measurement set using the classical ac bridge technique. In addition, through the use of small amplitude ac currents one is not faced with the problem of deviations from equilibria which may occur with large currents. Finally, this method, using the described instrumentation, leads to higher accuracy than that obtained by sinusoidal ac polarography.
ACKNOWLEDGMENT We acknowledge the loan of a lock-in analyzer from EG&G Princeton Applied Research. Registry No. Zn, 7440-66-6;KN03, 7757-79-1;KC1,7447-40-7; Zn(Hg), 11146-96-6;Hg, 7439-97-6;adenine, 73-24-5. LITERATURE CITED (1) Sluyter-Rehbach, M.; Sluyters, J. H. "Electrochemical Chemistry"; Marcel Dekker: New York, 1970; Voi. 4, pp 1-121. (2) Archer, W. I.; Armstrong, R. D. "Electrochemistry"; Burlington House: London, 1978; Vol. 7, pp 157-201. (3) Armstrong, R. D.; Bell, M. F.; Metcalfe, A. A. "Electrochemistry"; Burlington House: London, 1978; Vol. 6, pp 98-121. (4) Grahame, D. C. Chem. Rev. 1947, 4 7 , 441-501. (5) Grahame, D. C. J. Am. Chem. SOC.1941, 6 3 , 1207-1215. (6) Gabrlelli, C. "Identificatlon of Electrochemlcal Processes by Frequency Response Analysls"; Solartron: Paris, 1980. (7) Brleter, M. W. J. Electrochem. SOC.1965, 772, 845-849. (8) Tshernikovskl, N.; Glleadi, E. f/ectrochlm. Acta 1971, 76, 579-584. (9) Bower, 0. P.; Caldwell, I. flectrochim. Acta 1981, 26, 625-629. (10) Feller, H. G.; Ratzer-Schelbe, H. J.; Wendt, W. Electrochim. Acta 1972, 77, 187-195. (11) Cai, S. M.; Liu, C. Y.; Wilhelm, S. M.; Hackerman, N. Extended Abstract of Electrochemlcal Society 161th Meeting, No. 697, Montreal, 1982. (12) Operating and Service Manual for EG&G Prlnceton Applied Research Model 5204 Lock-in Analyzer, Prlnceton, NJ. (13) O'Haver, T. C. J. Chem. fduc. 1972, 498 A131-134 and A211-222. (14) Vetter, K., "Electrochemische Klnetik"; Sprlnger-Verlag: Germany, 1981. (15) Knoshila, H.; Chrlstian, S. D.; Dryhurst, G. J. Necfroanal. Chem. 1977, 83, 151-166. (16) Parsons, R. "Handbook of Electrochemlcal Constants"; Butterworths: Longon, 1959.
RECEIVED for review August 19, 1982. Accepted October 7, 1982. The support of the National Science Foundation (Grant CHE 7921536) and the National Institutes of Health (Grant GM 25172) is gratefully acknowledged.
Impedance Measurements for Evaluating the Stability of Aqueous Saturated Calomel Reference Electrodes in Nonaqueous Solvents Karl M. Kadish, * Sheng-Min Cal,' Tadeusz Malinski, Jlan-Quan Dlng, and Xiang-Oin Lin Department of Chemistry, University of Houston, Houston, Texas 77004
The increased use of nonaqueous solvents for electrochemical studies has led to the utilization of numerous novel reference electrodes suitable for these solvents. These electrodes On leave from the Chemistry Department, Peking University, Beijing, People's Republic of China.
and their specific advantages and disadvantages are discussed in several comprehensive monographs (1-4). In almost all cases the authors recommend against the use of an aqueous saturated calomel electrode (SCE) in nonaqueous media, However, despite these warnings, the most often utilized electrode in nonaqueous media remains as the saturated
0003-2700/83/0355-0163$01.50/00 1982 American Chemlcal Soclety
164
ANALYTICAL CHEMISTRY, VOL. 55, NO. 1, JANUARY 1983
Table I. Types of Saturated Calomel Electrodes Investigated electrode no. model type frit 1 IBM Model 4 3 porous vycor 2 IBM Model 4 3 porous vycor 3 “home made” porous vycor 4 Metrohm EA402 asbestos fiber 5 Metrohm EA402 asbestos fiber 6 Metrohm EA404 asbestos fiber I Fisher Dri-Pak asbestos fiber 8 Fisher Dri-Pak asbestos fiber 9 Radiometer K cellulose pulp calomel electrode. A survey of the literature in Inorganic Chemistry from 1979 to 1982 showed that over 75% of the 50 plus studies concerning some aspect of nonaqueous electrochemistry were carried out with an aqueous SCE. For measurements in nonaqueous media the SCE is usually connected by means of a nonaqueous salt bridge (e.g., nonaqueous solution of tetraalkylammonium salt) to the electrolyte under study. This kind of double junction bridge separates the investigated nonaqueous solution from contact with the aqueous KC1 solution in the SCE. However, the choice of this particular bridge arrangement in conjunction with an SCE is not perfect because potassium chloride has a limited solubility in many nonaqueous solvents ( 4 ) . In addition the junction can be readily clogged, which leads to erratic junction potentials due to nonaqueous contamination in the frit. In these cases there is also the possibility that traces of nonaqueous solvent cannot be easily removed because of adsorption, even if the SCE is stored in aqueous saturated KC1 for a long time after use. Special problems are also encountered in dichloromethane. This low dielectric constant solvent is not miscible with water and tends to promote clogging at the capillary tip. This leads to both erroneous potentials and high cell resistances which will produce distorted current voltage curves (due to iR loss). Because of these above problems the stability of an SCE after use in nonaqueous solvents must be checked frequently by comparison with standard reference electrodes which have not been subjected to use in nonaqueous media. This most often involves a check of the potential difference between the two electrodes which should not exceed 1-2 mV. For a number of years our laboratory has used the saturated calomel electrode as a reference electrode in nonaqueous media (5-7). For these studies we have utilized a bridge to separate the aqueous and the nonaqueous solvent systems. This bridge was filled with the nonaqueous solvent containing 0.1 M tetrabutylammonium perchlorate. As required, we have periodically checked the potential against a fresh electrode and, in addition, have reported potentials vs. a Fc+/Fc (ferrocene/ferrocenium) internal standard. While this method seemed to work well and helped us to identify faulty reference electrodes in most cases, it was evident that the measurement of potential alone was not sufficient to identify an electrode which was not operating “correctly”. For this reason we have devised a technique utilizing the ac impedance method which will give us the value of resistance and capacitance of an electrode after use in a nonaqueous solvent. In this paper we report the results of a typical measurement of reference electrode potential and resistance before and after use in nonaqueous solvents. Measurements were made in dimethyl sulfoxide (Me2SO), acetonitrile (CH&N), and methylene chloride (CH2C12),although results are only reported in this latter solvent, which is the worst case. The eight commercially available electrodes and one homemade electrode listed in Table I were investigated. All of the electrodes were used extensively in CH2Clzalthough three of these had
I
ll I I
w I
Flgure 1. Schematic diagram of instrument for ac impedance measurements: VCFO, voltage controlled frequency oscillator; lock-in, lock-In analyzer; C, cell; RS, reference signal input; TE, test electrode; CE, ac counterelectrode.
been discarded due to some previously undefined malfunction. Of the nine electrodes, three contained porous vycor frits, five contained asbestos fiber frits, and one contained a cellulose pulp frit. Our aim in the study was to report the potential, resistance, and capacitance (the last two were calculated from impedance measurements) of these electrodes a t a given point in time and to then select three of these electrodes for measurements after long-term storage in CHzClz(which might simulate an experiment). At the end of this time the electrodes would be placed in saturated KCl and the time to reach reequilibrium measured.
EXPERIMENTAL SECTION Reagents and Solutions. Three different nonaqueous, aprotic solvents were used throughout this study. CH2C12was obtained from Fisher Scientific as technical grade and was distilled from P205and stored in the dark over activated 4-A molecular sieves prior to use. MezSO (Eastman Chemicals) and CH&N (Coleman) were received as reagent grade from the manufacturer and were dried over 4-A molecular sieves prior to use. KC1 (Fisher Scientific) analytical grade was used as received. Instrumentation and Measurements. Impedance measurementa were made with a home built instrument whose design has been described in the literature (8,9). The block diagram of this instrument is shown in Figure 1. A large (2 cm2)platinum mesh electrode was used as an alternating current counterelectrode (CE). A Hewlett-Packard Model 3310 generator (VCFO) which generated a sinusoidal wave with a frequency of 1000 Hz was connected to the CE via a 2000-kQresistor and a 200-wF capacitor. The amplitude and phase of the signal voltage between the test electrode (TE), which in this case was the SCE, and the counterelectrode were measured with a PAR Model 5204 lock-in analyzer. For measurements of potential a digital voltammeter (Ballantine-STD) with high input resistance was used. All reported potentials are referred to an IBM Model 43 standard reference electrode which is listed as electrode number 1in Table I. RESULTS AND DISCUSSION Using a lock-in analyzer one can simultaneously measure the resistive component (ac voltage component VR)and capacitive component (ac voltage component Vc) of an electrochemical cell. The resistive component is in-phase with the ac current, i, and the capacitive component is i ~ / 2outof-phase with the current. A total resistance, R , can be calculated from eq l and the total capacitance, C, from eq 2 where f is the frequency of the alternating current.
Because a resistance of 2000 kQ was inserted into the circuit (see Figure 1) the ac current passing through the SCE is very small (less than 0.05 PA). Use of a small current is necessary in order to prevent polarization of the electrode during the
.
ANALYTICAL CHEMISTRY, VOL. 55, NO. 1, JANUARY 1983
Table 11. Properties of Saturated Calomel Electrodes potential differ - capacience,a tance, resistance, electrode frit mV pF n porous vycor 40.0 400 40.0 250 -0.1 porous vycor 0.0 9.6 380 porous vycor ‘11.0 850 -0.3 asbestos fiber 1000 -0.3 :14.0 asbestos fiber 9.2 760 asbestos fiberb -1.2 -0.4 4.0 1500 asbestos fiber 0.1 3.0 1660 asbestos fiber 1.9 280 -0.9 cellulose pulp a Defined as potential difference between given reference electrode and electrode number 1. Due to some undefined probleims these electrodes had not been used in experiments for periods of 6 moaths or longer. The values presented are after treatment of the electrode as defined in the text. The original values are given in Table 111.
165
P-
co
Flgure 2. Equivalent circult of t h e electrochemical impedance for reference electrode: R e , electrolyte resistance; R,,, charge transfer resistance; C,,double layer capaclty; R,, diffusion resistance (Warburg resistance);C,, diffusion capacity (Warburg capacity).
measurement process. In this study potentials were found to vary by less than 0.1 mV when 0.05 pA current was passed through the SCE for a period of up to 24 h. For purposes of our ineasurements the calomel reference electrode can be represented by the equivalent circuit shown in Figure 2. Because of the small current passing through the circuit, and the small solubility of Hg2C12,the reaction Hgz2+can be considered as negligible. Under these conditions the Warburg impedance (R, + C,) reaches a high value. Thus, the faradaic impedance contributions from Ret, R,, and C, are negligible as a first approximation. In this case the calculated impedance values of R and IC are approximately equal to Re and Cd, where Re represents the sum of solution resistance in the bridge and the SCE. (There is also a small
resistance of the connecting wires which is considered negligible.) The measured capacitance, Cd, is dependent on the capacitance of the double layer between the calomel, KCl, and mercury of the SCE, and gives important information about the contact area between the calomel paste and the mercury. Table I lists the nine electrbdes which were investigated in this study. These electrodes were from five different sources (four were commercially available and one was home made) and contained three different types of frits (porous vycor, asbestos fiber, and cellulose pulp). All of the electrodes had been used extensively in a variety of nonaqueous solvents. Table I1 lists the potential differences, capacitances, and resistances of the nine different electrodes listed in Table I. Six of these electrodes were in regular use before the measurement had been carried out and had been properly stored in saturated KCl. Three of the electrodes (no. 2, 6, and 8) had been neglected due to some undefined problem which involved either unstable potential or abnormally high resistance. As might be expected, the three electrodes that had been discarded showed large deviations in potential, capacitance, or resistance. In these cases the values presented in Table I1 are after appropriate treatment of the electrode. Electrode 2 had a dirty frit and was not saturated with KC1; electrode 6 had a poor connection between the mercury and
Table 111. Properties of “Bad” Calomel Electrodes Before and After Treatment potential difference,b capacitance, resistance, electrode problem frit treatment a mV PF n 2 c porous vycor before -14.4 13.0 400 after -0.1 40.0f 250 6 cl asbestos fiber before -36.6 0.3 8600 after -1.2 9.2f 760 8 I3 asbestos fiber before -1.5 2.5 3000 after 0.1 3.0f 1660 a According to the procedure described in the text. Potential differences measured vs. SCE labeled as number 1 (see Table I). Dirty frit; KC1 solution was not saturated. Poor connection between mercury and calomel paste. e Frit was dirty and clogged. f ABter improving the contact between mercury and calomel. Table IV. Change of Potential and Resistance of Electrodes After Storage in CH,Cl, and Reequilibrium After Saturated KC1 Storage electrode 2 electrode 4 electrode 9 time, -__ solvent min A E , b mV R, A E , mV R , C2 AR, F E , mV R, AR, n CH,Cl, 0 0.0 380 0 0.0 850 0 0.0 280 0
saturated KCla
120 480 720 1440 1 2 3 4 10 30
0.0 0.0 0.1 0.3 0.3 0.2 0.1 0.0 0.0 0.0
400 400
380
Time measured after 24 h of storage in CH,Cl,. a function of time. a
20 20
0
0.0 -0.2 --0.2 -0.4 -0.4 -0.3 -0.3 --0.2 -0.1 0.0
0.0 0.2 0.3 9 00 900
50 50
850
0
Change in potential as a function of time.
0.5
0.5 0.4 0.3 0.2 0.1 0.0
320 320
40 40
280
0
Change of resistance as
Anal. Chem. 1983, 5 5 , 166-169
166
the calomel paste; and electrode 8 had a frit which was clogged and dirty. The properties of these three electrodes before and after treatment are listed in Table 111. As part of the treatment every frit was washed and cleaned, the connection between the mercury and the calomel paste was fixed, and the electrodes were filled with freshly prepared saturated KCl. These electrodes were then stored for 48 h in saturated KC1 and measurements repeated. In every case the resistance was lower and the capacitance was higher. In addition, the potential of all three electrodes was much closer to the potential of our standard reference SCE labeled as number 1. The data in Table I1 show that all of the electrodes have similar correct potentials. Thus, this might be one criterion for stating all of the electrodes are good for use as standard reference electrodes. On the other hand, the electrodes differ substantially in terms of their capacitance and resistance and can be grouped according to both the model of the electrode and the type of frit. The lowest resistances are found for electrodes with porous vycor and the cellulose pulp frits (250-400 R) while the highest are for the electrodes with asbestos fiber frits (760-1660 Q). In this grouping the lower values are found for the three Metrohm electrodes while the highest are for the two Fisher electrodes. As seen in Table I1 the values of capacitance vary between a low of 1.9 and 40 pF. These values reflect the contact between the mercury and the calomel which is an important factor in the stability of the electrode potential. The higher the value of capacitance, the larger the contact area, and the more stable will be the electrode potential. All of the electrodes from a given manufacturer had similar values of capacity as might be expected. The IBM electrodes (1 and 2) had identical values of 40 pF and were larger than the three Metrohm electrodes (4, 5, and 6) which ranged from 9.2 to 14.0 pF. The home made electrode (electrode 3) had a capacity of 9.6 pF. This electrode had a relatively small dispersion of mercury in the calomel paste and also had a shorter KCl bridge than did electrode 1 and 2. Finally, both the Fisher and the Radiometer electrodes had a low capacity of 1.9-4.0 pF. Data collected in Table IV represent the influence of CH2ClZ on the potential and resistance of three representative electrodes. Electrode 2 contains a porous vycor frit, electrode 4 an asbestos fiber frit, and electrode 9 a cellulose pulp frit. Table IV also includes data on reequilibrium of the electrodes in KC1 after storage in CH2Clzfor periods of 2-24 h.
After each interval of time the electrodes were transferred to an aqueous saturated KCl solution and impedance measurements were made. Although only 30 min was required for reequilibrium in KCl (see following sections) the electrodes were stored in KC1 for a minimum of 3 h between measurements in CH2C12. After 2 h of storage in CH2C12,no change in potential was observed, as seen in Table IV. In fact, very little change was observed even after 24 h. In the very worst case only a 0.5 mV difference in potential and a 50 R increase in resistance were obtained. Similar small changes were observed in Me2S0 and CH3CN. More surprisingly immersion of the electrodes in saturated KC1 for 4-30 min after being in CH2C12for 24 h produces identical potentials and resistances as in the original measurement. In conclusion, one can see that potentials alone are not sufficient to characterize a properly operating reference electrode. Furthermore, the influence of nonaqueous solvents on an aqueous calomel reference electrode seems to be far less than had been expected, as any observed differences in potential and resistance caused by storing the electrodes in nonaqueous solvent for up to 24 h appear to be transient. This should not be construed as a recommendation for use of an aqueous SCE in nonaqueous solvents but should only be used as a guide in determining when an electrode might be faulty.
LITERATURE CITED Hills, G.J.; Ives, D. J. G. "Reference Electrodes"; Ives, D. J. G., Jane, G.J., Eds.; Academic Press: New York, 1961; Chapter 10. Strehlow, H. I n "The Chemistry of Nonaqueous Solvents"; Lagowski, J. J. Ed., Academic Press: New York, 1967; Vol. 11, Chapter 4. Butler, J. N. I n "Advances in Electrochemistry and Electrochemical Engineering"; Delahay, P., Ed.; Interscience: New York, 1970; Vol. 7, pp 106-114. Sawyer, D. T.; Roberts, J. L., Jr. "Experimental Electrochemistry for Chemists"; Wlley: New York, 1974. Kadish, K. M.; Morrison, M. M. J. Am. Chem. SOC. 1976, 98, 3326. Kadish, K. M.; Bottomley, L. A.; Brace, J.; Winograd, N. J. Am. Chem. SOC. 1980, 102, 4341. Bottomley, L. A.; Kadish, K. M. Inorg. Chem. 1981, 2 0 , 1348. Cai, S. M.; Liu, C. Y.; Wilhelm, W. M.; Hackerman, N. Extended Abstract of Electrochemical Society of l6lst Meeting, No. 697, Montreal, 1982. . Cai, S. M.; Malinski, T.; Lin, X.-G.; Ding, J. 0.; Kadish, K. M. Anal. Chem. 1983, 55, 161-163. ~~
RECEIVED for review July 23, 1982. Accepted September 24, 1982. This work was supported by the National Science Foundation (Grant CHE 7921536) and the National Institutes of Health (Grant GM 25172).
Back-Extraction with Three Aqueous Stripping Systems for 16 Elements from Organometallic-Halide Extracts J. Robert Clark" and John G. Wets U.S.
Geological Survey, Denver, Colorado 80225
Detailed stripping curves have been determined for Cu, Ag, Au, Zn, Cd, Hg, Ga, In, T1, Sn, Pb, As, Sb, Bi, Se, and Te, which are extracted by the Methyl isobutyl ketone-Amine synerGistic Iodine Complex (MAGIC) extraction system ( I ) . Stripping was accomplished with a HN03-H202 system, a CH3COOH-H2OZsystem, and a H2SO4--H2Oz system (2). The mechanisms by which stripping is accomplished include poisoning of amine ion exchange agents with anions that are
incompatible with the extraction system, and oxidation of halide (mostly iodide) complexing ions. Most of these 16 elements can be separated from many other elements in the organic extracts by sequentially stripping the organic phase with various combinations of these three systems. The MAGIC extraction system (1) makes it possible to concentrate and separate 18 trace elements (Pt and Pd, in addition to the list above) from analytically interfering rock
This article not subject to U S . Copyright. Published 1982 by the American Chemical Society
